Scholarlink

Background: Cyber threats pose an increasingly significant risk to software ecosystems worldwide, resulting in a pressing need for robust defense mechanisms to safeguard our digital ecosystems. Identifying and mapping the continuously changing cyber threat landscape is a means to mitigate their impact and protect digital assets correctly. Aim: We aim to explore the potential assessment value of data from publicly available bodies of knowledge in identifying global cyber security trends by applying topic modeling on publicly available data. Method: Mining and analyzing data from publicly available bodies of knowledge, such as popular developer platforms, offers a unique opportunity to explore and anticipate emerging cybersecurity trends worldwide. We analyzed Stack Overflow discussions using Latent Dirichlet Allocation, which were cross-referenced with the most common security vulnerabilities to verify accuracy. Results: Our results show that combining text mining and topic modeling can identify and categorize large-scale cybersecurity threats. We found twelve unique cyber risk topics, seven of which match the OWASP Top 10 Security Risks in 2021. This observation highlights the effectiveness of our approach in tracking major cyber threats and its potential as a robust cybersecurity analytics tool.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Canary in the Coal Mine: Identifying Cyber Threat Trends Through Topic Mining – Stack Overflow Case Study

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Matthijs Koier,
  • Hugo Kuipers,
  • Marre Slikker,
  • Slinger Jansen,
  • Katsiaryna Labunets

摘要

Background: Cyber threats pose an increasingly significant risk to software ecosystems worldwide, resulting in a pressing need for robust defense mechanisms to safeguard our digital ecosystems. Identifying and mapping the continuously changing cyber threat landscape is a means to mitigate their impact and protect digital assets correctly. Aim: We aim to explore the potential assessment value of data from publicly available bodies of knowledge in identifying global cyber security trends by applying topic modeling on publicly available data. Method: Mining and analyzing data from publicly available bodies of knowledge, such as popular developer platforms, offers a unique opportunity to explore and anticipate emerging cybersecurity trends worldwide. We analyzed Stack Overflow discussions using Latent Dirichlet Allocation, which were cross-referenced with the most common security vulnerabilities to verify accuracy. Results: Our results show that combining text mining and topic modeling can identify and categorize large-scale cybersecurity threats. We found twelve unique cyber risk topics, seven of which match the OWASP Top 10 Security Risks in 2021. This observation highlights the effectiveness of our approach in tracking major cyber threats and its potential as a robust cybersecurity analytics tool.